Architecture-based force-velocity models of load-moving skeletal muscles.

A predictive model of muscle force-velocity relationships is presented based on functional architectural variables. The parameters of Hill's equation describing muscle force-velocity relationship of nine muscles were estimated by their relationships with variables extracted from the whole-muscle length-force relationship and the percentage of slow-twitch fibres. Specifically, the maximal unloaded velocity (Vo) was estimated through multiple linear regression, from each muscle's fibre composition and the shortening range through which each muscle could produce active force. The maximal isometric force (Po) was also extracted from each muscle's length-force relationship. The ratio of Hill's dynamic constanta to Po and b to Vo, which determines the degree of curvature of the relation, was determined solely by the percent of slow-twitch fibres. This model was verified by fitting it to experimental force-velocity curves of nine different muscles in the cat's hindlimb. It was found that reasonable fits of force-velocity curves would be obtained with correlation coefficient in the range of 0.61 to 0.92, with an average of 0.82. The model predicted that muscles with relatively long shortening ranges would achieve higher maximal velocity, and that muscles with higher percentage of slow-twitch fibres had less pronounced curvature and lower maximal velocity in their force-velocity relationships. RELEVANCE: The results have direct implications in the design of neuroprosthetic limb control systems, which use electrical stimulation to restore function to muscles paralysed from spinal cord injury. The designer is enabled to optimally calibrate the controller according to the predicted individual force-velocity curves of different muscles by using the length-tension curves and fibre composition data available in the literature.